if you can't tell the boundaries, then it ain't property - research on

Bessen and Meurer Chapter 3 - 1

If you can’t tell the boundaries, then it ain’t property

1 Introduction

A successful property system establishes clear, easily-determined rights. Claritypromotes efficiency because “strangers” to a property can avoid trespass and otherviolations of property rights, and, when desirable, negotiate permission to use theproperty. The concepts in the last sentence are critical to understanding the performanceof the patent system. As we shall see, increasingly, patents fail to provide clear notice ofthe scope of patent rights. Thus, innovators find it increasingly difficult to determinewhether a technology will infringe anyone’s patents, giving rise to inadvertentinfringement. Similarly, they find it increasingly costly to find and negotiate thenecessary patent licenses in advance of their technology development and adoptiondecisions. Thus, clearance procedures that work well for tangible property are undercutby a profusion of fuzzy patent rights.

An ideal patent system features rights that are defined as clearly as the fence arounda piece of land. Realistically, no patent system could achieve such precision, but ourcurrent system appears to be critically deficient in this regard. The comparison to tangibleproperty is informative. In the last chapter, we suggested that patent law shares manydoctrinal features with the law of tangible property, but that the application of thesedoctrines might be substantially different.

This chapter compares in detail the law and institutions that promote patent noticewith the corresponding law and institutions that provide notice for tangible property. Thepatent system fares badly in this comparison; certain institutions that contribute to clearnotice are pitifully underdeveloped. It is hardly surprising, then, that patents, unliketangible property, have a significant problem with inadvertent infringement. Moreover,we find some evidence that notice problems have been getting worse.

2 Bad Fences Make Bad Patents

2.1 Two Kinds of Inadvertent Infringement

We begin by looking in detail at inadvertent infringement, notice and clearance.Inadvertent patent infringement often arises either when a firm independently invents atechnology that was previously patented, or when a firm attempts to design non-infringing technology that competes with a patented technology. Disputes arise in somecases because alleged infringers are not aware of the earlier invention and the purportedpatent rights. Other disputes arise because the set of potentially relevant patents is large,the scope of the claims is vague, and many of the claims may be invalid. Under theseconditions, designing around patents is difficult and clearing the rights may beprohibitively expensive.

The BlackBerry case illustrates the first type of inadvertent infringement. RIM,the maker of the BlackBerry, was ensnared in a long-running patent infringement lawsuit


with a company called NTP.1 NTP co-founder Thomas Campana Jr. tried and failed tocommercialize wireless e-mail but he did acquire several patents relating to thetechnology. Mike Lazaridis, the founder of RIM, independently invented similartechnology, which he patented, and turned into the BlackBerry without gettingpermission to use Campana’s patents. After five years of litigation, and facing the threatof injunction, RIM settled with NTP and agreed to pay $612.5 million.

The momentous dispute between Kodak and Polaroid nicely illustrates the secondtype of inadvertent infringement.2 Kodak was the dominant firm in Americanphotography from the earliest days of the industry. Newcomer Polaroid introduced instantphotography in the late 1940s with help from Kodak; for several years they had nosignificant competition in the U.S. market. Kodak explored instant photography and evenproduced an instant film technology that was used on the Lunar Orbiter in 1966, but theydelayed entry into the consumer market until 1976. One factor that slowed Kodak’s entrywas the great care they took to invent around Polaroid’s patents. Kodak started researchin 1969 on an instant photography product that would compete with Polaroid. From thebeginning they consulted with patent lawyers to make sure they steered clear of Polaroidpatents. Kodak believed its technology was very different from Polaroid’s. Kodak’sformer senior vice president and general counsel Cecil Quillen Jr. believed: “The Kodakchemistry worked exactly backward from the Polaroid chemistry….” Nevertheless,Kodak lost the patent suit that Polaroid filed one week after Kodak entered the market.They paid Polaroid about $900 million and exited the instant photography market.

2.2 Patent Clearance

Of course, RIM and Kodak could have avoided these problems if they hadlicensed the patented technology up front (had the other parties been willing to license).That is, they could have cleared the rights in advance. This sort of thing happensroutinely with tangible property and even with other intellectual property such ascopyrights.

The world of movie production and copyright clearance provides a glimpse ofwhat the patent system should aspire to achieve in terms of notice and clearance. Moviesoften incorporate a variety of pre-existing copyrighted works. They contain pre-existingsound recordings, new versions of old songs, special effects video, art work displayed insets, and dialog based on screenplays, and sometimes plays, novels, and other movies. Amovie producer must be careful to obtain copyright permission from the owners of thecopyrights on these other works. If a producer fails to clear the rights, even to a work thatplays quite a minor role in the movie, the owner of the infringed copyright can stop allperformance of the movie.

Experienced Hollywood movie producers manage copyright clearance withoutmuch trouble. The transactions to acquire the necessary rights are relativelystandardized.3 The producer usually gets a license to use pre-existing art, music, video,

1 Barrie McKenna, Paul Waldie and Simon Avery, Patently Absurd, Globe and Mail January 28, 20062 Many commentators credit this case with opening eyes in the business world to the growing danger

from patents after the creation of the Federal Circuit. For a behind the scene discussion of the case seeChapter 3 in Fred Warshofsky, Patent Wars: The Battle to Own the World’s Technology (Wiley, 1994).

3 The leading copyright treatise has a chapter on motions pictures that starts with a rights clearance


and text. Copyrights to copyrighted work created during the production of the movie areusually assigned to the movie producer through contracts signed by the manycollaborators who contribute to the movie.

Motion picture producers and technology developers share comparably complexintellectual property environments. But technology developers have far more difficultynegotiating clearance contracts, or steering clear of problematic patents. The BlackBerryand instant photography sagas are two of many recent examples of sensational failures ofpatent notice and clearance. Let’s try to understand why these failures occurred.

Why didn’t RIM clear patent rights in advance of the BlackBerry introduction theway movie producers clear copyrights in advance of movie distribution? One reason isthat Lazaridis did not build on Campana’s technology the way a movie screenplay buildson the plot of a novel. Campana and Lazaridis independently invented various aspects ofthe technology, likely without knowing their future rival existed.4 RIM first learned aboutNTP and Campana’s patents in early 2000 when NTP sent letters to several companies,including RIM, warning them about NTP’s wireless e-mail patents. This was ten yearsafter RIM started developing wireless technology, four years after RIM introduced itsprototype of the BlackBerry, and two years after RIM signed contracts with Canadian andAmerican phone companies to supply wireless e-mail service.

But surely, a sophisticated business like RIM was aware that independentinvention is not a defense to patent infringement. In fact, an innovator can even get apatent on his technology (as Lazaridis did), and still be liable for infringement ofsomeone else’s patent. Why didn’t RIM search for patents that they might infringe? Afterfinding such patents they could have negotiated a license, or redesigned their technologyto avoid infringement. We don’t know why RIM didn’t find and deal with Campana’spatents at an earlier date, but we can make an educated guess. If RIM had searched, theywould have found many patents granted to many inventors with uncertain scope andvalidity that might apply to various aspects of BlackBerry. Further, RIM would still haveto worry about unpublished applications that might mature into patents covering theirtechnology. We suspect that often, it is simply bad business to search for patents andnegotiate licenses in advance of technology adoption. The costs of sorting through a largenumber of uncertain property rights is larger than the expected cost caused when any onepatent is asserted against the innovator. We cannot be sure, but we would not be surprisedif RIM’s failure to do a patent search was, at the time, the best decision.

Kodak chose the opposite path, but fared no better. They retained a leading patentexpert named Carr to advise them how to avoid patent infringement. Carr conducted anexhaustive review of potentially relevant patents and Kodak carefully developed

checklist, and contains a variety of forms for standard contracts. Nimmer and Nimmer §25.4 Campana developed pager inventions in the early 1980s and started working on pagers with text

messages around 1985. The trial court judge noted: “It is apparent that RIM developed and conceivedits BlackBerry products entirely independent of the Campana patents.” NTP, Inc. v. Research inMotion, Ltd., 270 F. Supp. 2d 751, 755 (D. Va. 2003). A decade before Campana, Geoff Goodfellowinvented wireless e-mail and published the idea on an Arpanet mailing list. Like Campana, his start-upfailed. He didn’t patent the idea because he thought it was obvious. RIM identified three other inventorswho may have predated Campana. John Markoff, In Silicon Valley, a Man Without a Patent, NYTimesApr. 16, 2006.


technology they thought was outside of Polaroid’s patent fences.5 Nevertheless, Polaroidasserted that Kodak had “willfully” (knowingly) infringed its patents, but the trial judgerefused to award enhanced damages for willfulness and concluded: “The uncontrovertedfacts demonstrate that Kodak consulted Mr. Carr early and often as it developed itsinstant integral photography system.”6

Kodak’s patent review was thoroughly integrated with its R&D on consumerinstant photography. The review started seven years before a commercial product wasintroduced. Carr studied more than 250 patents, many owned by Polaroid and manyowned by others. He “rendered 67 written and countless oral opinions on both the filmand camera patents.” In the lawsuit, Polaroid focused on thirty-four claims contained inten of their patents. Three of the patents were invalidated. Kodak was held liable forinfringing twenty claims in the remaining seven patents.

Carr had advised Kodak that the claims in those seven patents were either invalidor not infringed. “Over the course of three years Mr. Carr reviewed more than 50potential imaging chemistries for Kodak. Eventually, after working closely with Mr. Carrand performing tests he requested to make certain he understood how the chemistryworked,” Kodak chose an imaging technology that was later ruled infringing.�Kodak alsoincorporated design changes to mechanical features of its camera that Carr thought wouldkeep them clear of Polaroid patents. The trial judge concluded that Carr “discussed theKodak technology with Kodak engineers, and gave his considered advice….That advicesimply turned out to be wrong.”

These two examples show that the patent notice function breaks down for twosorts of reasons. First, a technology investor may not be able to unambiguously determinethe scope and validity of a set of patents, as in the Kodak example. Second, even if scopeand validity could be determined with reasonable accuracy, it may be too costly to do so.Below we look at the law and institutions that give rise to both of these situations.

2.3 Land versus Technology

The BlackBerry and instant camera cases illustrate how notice and clearancefailure cause expensive patent disputes. Regrettably, patent notice and clearanceproblems are widespread. One good way to recognize the size of the problem is bycomparing property rights in land to patents.

Real property law gives land owners a clear view of property boundaries. Beforebuilding a costly structure near the perimeter of one’s land, a sensible land owner willconsult a lawyer and conduct a survey of the land. Land records, a survey, and a titlesearch give a potential builder clear notice of where her property ends and her neighbor’sbegins. We rarely hear about lawsuits caused because someone inadvertently built astructure on, or made some other investment in another’s property.7

5 The failure of Kodak to negotiate a license from Polaroid might be explained by their divergent beliefsabout the scope and validity of the relevant Polaroid patents.

6 Polaroid Corp. v. Eastman Kodak Co., 1990 U.S. Dist. LEXIS 17968 (D. Mass. 1990) at 236.7 Over the past three years there have been only four lawsuits in California concerning good faith

improvement of land. The disputes involved one party intruding onto another’s property and making aninvestment that delivered value to the other’s property. The investments involved tree cutting, a road,crops, and landscaping. California law defines a good faith improver as: “A person who makes an


A prospective land buyer will take the same steps before buying land. Normally,any clouds on the title to the land will be cleared before the sale is completed. Theprocess of examining property rights to land is routinely provided by a robust market thatcombines title examination with title insurance.

In contrast, patent search and clearance is hardly routine. The process is costlyand inconclusive. Typically, the risk of infringement that remains after a competentpatent review is so unpredictable that it is virtually uninsurable.8 Similarly, uncertaintyabout scope and validity undermine the market for patent enforcement insurance.

Overall, the legal infrastructure supporting patent law performs poorly comparedto the infrastructure supporting real property law. A few numerical comparisons lay barethe stark contrast. Title insurance on a $150,000 mortgage runs about $450 in Iowa, astate with a relatively efficient system.9 The revenue from these policies is more thanadequate for the insurers to cover their cost of searching and examining the title,defending against future lawsuits, and paying losses that result from lawsuits.Impressively, search and examination is so effective that insurers nationwide pay onlyabout 5 percent of premium dollars on claims.10

Now imagine that an innovative firm purchases insurance against patent lawsuitsthat might arise after it adopts a new technology. What would it cost an insurancecompany to search and examine patent databases, pay to defend patent lawsuits, and payfor damages (and perhaps even the costs from an injunction)? We do not have an estimateof the search and examinations costs, but unlike title insurance, the lion’s share of thecost would probably arise from lawsuits. We estimate in Chapter 6 that the expected costof defending patent lawsuits is now about 19% of the cost of R&D investment. This ratiois a lower bound on the cost of our hypothetical infringement insurance. The cost ofsearch and examination plus the expected cost of defending title lawsuits is much less.Dividing $450 by $150,000 gives a ratio of 0.3%, which is less than one sixtieth the ratiofor patents.

A similar comparison between patents and copyrights would probably be equallydiscouraging, but we are not aware of any data on either copyright infringementinsurance or copyright litigation costs. Two pieces of information suggest that copyrightdoes far better than patent in providing notice and facilitating clearance. First, copyrightclearance is relatively standardized in the movie, music, and publishing industries.Second, copyright infringement insurance is available to protect against risks created byerrors or omissions in the clearance process.11

improvement to land in good faith and under the erroneous belief, because of a mistake of law or fact,that he is the owner of the land.” (Code Civ. Proc., § 871.1.)

8 Richard S. Betterley, The Betterly Report, 2006, available at: Richard S. Betterley,http://www.betterley.com/adobe/ipims_06_nt.pdf

9 Melynda Dovel Wilcox, Cut Your Title Insurance Costs, December 2005,http://www.kiplinger.com/personalfinance/basics/managing/insurance/title.html. In other states the costmay be as high as $1500. The $450 figure, at the low end of the range, is probably the best measure ofthe social costs of the land title system. There are complaints and recent government investigations ofanti-competitive practices in the title insurance industry that drive up premium prices.

10 M. Martin Boyer and Charles M. Nyce, Market Growth, Barriers to Entry, and Banks as InsuranceReferral Agents: Evidence from the Title Insurance Industry, February 2005.11 For example, Chubb provides Reputation Injury and Communications Liability covers costs from

copyright and trademark infringement lawsuits. http://www.chubb.com/businesses/cci/chubb1105.pdf


The software industry is especially interesting because patent and copyrightlawsuits are both possible. The IT Compliance Institute notes that sound IT policy canavoid most copyright infringement and trade secret misappropriation, but patentinfringement is virtually impossible to avoid; insurance against copyright infringement isaffordable, but patent infringement insurance is prohibitively expensive for mostcompanies.12

2.4 Differences in Law and Institutions

The costs of insurance and clearance are much lower for tangible property thanfor patents. Why? There are sharp differences between patents and tangible propertyregarding the law and institutions that promote clear notice. Consider the followingcontrasts between patents and land:

1. Fuzzy and unpredictable boundaries. Land can be inexpensively surveyed and thesurvey boundaries carry legal weight. While surveyors can plainly map the wordsin a deed to a physical boundary, it is much harder to map the words in a patent totechnologies. Not only are the words that lawyers use sometimes vague, but therules for interpreting the words are also sometimes unpredictable. In contrast,although innovators can obtain expensive legal opinions about the boundaries ofpatents, these opinions are unreliable. There is no reliable way of determiningpatent boundaries short of litigation.

2. Public access to boundary information. The documents used to determineboundaries for both land and patents are publicly available. However, it is possiblefor patent owners to hide the claim language that defines patent boundaries frompublic view for many years and this is being done with increasing frequency.

3. Possession and the scope of rights. Generally, tangible property rights are linkedclosely to possession, hence, the classic phrase: possession is nine points of thelaw. Patent law also requires possession of an invention, but often this requirementis not rigorously enforced. Consequently, courts sometimes grant patent ownersrights over technology that is new, different, and distant from anything theyactually made or possessed. Not surprisingly, this practice makes patentboundaries especially unclear in fast-paced technologies such as biotech andcomputer software.

4. The patent flood. Clearance costs are affected by the number of prospective rightsthat must be checked for possible infringement. Investments related to a new landuse rarely implicate very many parcels of land. Also, we noted in Chapter 2 thatproperty systems include features that discourage fragmentation of land rights.Investments related to new technology frequently need to be checked against largenumbers of patents. Although the patent system has features that discourage patentproliferation (notably the non-obviousness requirement), empirical evidencesuggests these are not working well.

These four differences in combination can explain the sharp difference in theperformance of the notice function and the corresponding differences in inadvertent

12 http://www.itcinstitute.com/display.aspx?id=160


infringement. We explore them in more detail over the remainder of this chapter.

3 Fuzzy and unpredictable boundaries

3.1 Institutional Failure: Why the Notice Function Works for Property but not forPatents

Property law provides good notice to potential purchasers about the propertyrights relevant to a contemplated new land use. Patent law often fails to provide goodnotice to innovators about the patent rights relevant to adoption of a new technology.Property law has stable doctrine and flourishing institutions designed to transmit clearnotice. Patent law lacks both.

Suppose a real estate investor wants to acquire a certain parcel of land and build amall. Her lawyers should be able to tell her that she will need to acquire rights to certainparcels A and B, but not nearby parcels C and D. They should also be able to tell herwhether there are any mortgages, easements, future interests, or other interests in therelevant property. To a great extent, public documents give notice of these interests andprovide evidence of their validity or invalidity.

The work of real estate lawyers is complemented by the work of title agents andsurveyors. Both are licensed professionals who contribute in important ways to thesuccess of property law in providing clear notice. Surveyors work from the plotdescription found in the deed or in a registered survey and, using well-recognizedstandardized methods, plot out the boundary lines. In most states, surveyors’ boundariesare legally recognized. A typical survey for a residential property costs several hundreddollars. Title agents search deeds and other public information to help lawyers judge thevalidity of property titles. A title search for a residential property with title insurance alsotypically costs several hundred dollars.

The patent system also has licensed patent agents and patent lawyers. They canalso render a judgment on whether a proposed investment (technology) infringes on validpatents owned by others. They judge both the boundaries and the validity of thesepatents. They also work by reading legal documents—issued patents—and applying themto the matter at hand. However, there are at least two important differences. First, a legal“opinion letter” on a technology typically costs about a hundred times more ($20,000 to$100,000). Second, this finding carries little legal weight.13

Indeed, no one except appeals court judges seem to be able to provide a definitiveanswer on potential infringement and they often disagree (Wagner and Petherbridge2004). It is true that patent examiners determine the boundaries of each patent—this isnecessary for them to determine whether the patent meets the criteria of patentability,e.g., whether the patent is truly novel. But patent examiners don’t record theirinterpretation of the boundaries of the patent and, even though courts presume that theirdecisions about patent validity are correct, courts pay little heed to the boundaries thatpatent examiners use to make these determinations.

13 Obtaining an opinion letter reduces the likelihood that an alleged infringer will be found to have“willfully” infringed (Knorr-Bremse Systeme Fuer Nutzfahrzeuge Gmbh v. Dana Corp.), however,opinion letters do not accurately predict whether infringement will be found, as in the Kodak case.


District courts hold hearings to interpret the boundaries of patents in lawsuits. Insome cases they call on expert witnesses to provide understanding of the technology andindustry usage of technical terms. But these determinations get no deference at theappellate level, either. In fact, the Federal Circuit reverses the district court judge’sconstruction of 34.5% of the claim terms appealed and this percentage has increased overthe course of the 1990s.14

These difficulties arise, of course, because it is much more complicated to mapthe boundaries of a technology from a verbal description than it is to map a plot of landusing a standardized surveyor’s description. This problem arises partly from the nature ofthe subject matter and partly from the difference in techniques used to do the mapping.We look at the law of patent claim interpretation in the next section.

But the hard fact is, innovators cannot quickly and easily obtain a reliablejudgment on whether prospective technology infringes on others’ patents. Perhaps in anearlier time, when technology was simpler, this was not such a serious problem becausethe ambiguity of patent claims was not so great. But as we develop below, there arereasons to think that this ambiguity has been increasing substantially in recent years. Inaddition, changes made during the 1990s in the legal methods used to determine theboundaries of patents appear to have made the uncertainty even greater.

3.2 A Nose of Wax?

Patent documents are typically long and obscure. The most obscure and mostimportant part of each patent is the set of claims found at the end of the document. Patentclaims create property rights. Each claim is a single sentence — an odd sort of sentenceto be sure, which might run on for several paragraphs. Collectively, they determine thescope of the owner’s right to exclude — they are the fences that mark the inventor’sproperty. Giles Rich, the most famous patent judge of modern times observed: “[T]hemain purpose of [patent] examination… is to try to make sure that what each claimdefines is patentable. To coin a phrase, the name of the game is the claim.”15

The game is stacked in favor of inventors and against examiners and the public.Examiners get an average of 18 hours to read and understand the application and makesure that each claim is valid. Patent law imposes many validity requirements. Timepressure means that inventors will often be able to push through questionable claims.Critics of the patent system quite properly complain that the Patent Office frequentlyapproves of claims that are obvious or lack novelty.

We want to begin by highlighting a different problem that has not attracted muchcomment — the issuance of vague claims. In order to be valid, the claims must meet therequirement found in paragraph 2, of section 112: “claims particularly pointing out anddistinctly claiming the subject matter which the applicant regards as his invention.” TheSupreme Court long ago recognized that “[t]he claim is a statutory requirement,prescribed for the very purpose of making the patentee define precisely what his

14 Kimberly A. Moore, Markman Eight Years Later: Is Claim Construction More Predictable? 9 Lewis &Clark L. Rev. 231 (2005).

15 Giles Sutherland Rich, Extent of Protection and Interpretation of Claims-American Perspectives, 21Int’l Rev. Indus. Prop. & Copyright L. 497, 499 (1990).


invention is; and it is unjust to the public, as well as an evasion of the law, to construe itin a manner different from the plain import of its terms�”16

�Nevertheless, patent applicantssometimes game the system by drafting ambiguous patent claims that can be readnarrowly during examination so they avoid a novelty rejection, and broadly duringlitigation which supports a finding of infringement. The Freeny patent discussed intChapter 1 is one example. Limited resources mean that the Patent Office does a poor jobof monitoring the clarity of patent claims, and thus notice suffers.

The Federal Circuit, the appeals court that sets most patent law standards, has notpromulgated rules to restrict vague claim language effectively. In fact, the court itself isreluctant to invalidate an indefinite claim. Allison and Lemley (1998) find that only 5.8%of invalidations are based on claim indefiniteness. The Federal Circuit explained:

We engage in claim construction every day, and cases frequently present close questionsof claim construction on which expert witnesses, trial courts, and even the judges of thiscourt may disagree. Under a broad concept of indefiniteness, all but the clearest claimconstruction issues could be regarded as giving rise to invalidating indefiniteness in theclaims at issue. But we have not adopted that approach to the law of indefiniteness. Wehave not insisted that claims be plain on their face in order to avoid condemnation forindefiniteness; rather, what we have asked is that the claims be amenable to construction,however difficult that task may be. If a claim is insolubly ambiguous, and no narrowingconstruction can properly be adopted, we have held the claim indefinite. If the meaningof the claim is discernible, even though the task may be formidable and the conclusionmay be one over which reasonable persons will disagree, we have held the claimsufficiently clear to avoid invalidity on indefiniteness grounds. Exxon Research andEngineering Co. v. U.S., 265 F.3d 1371, 1375 (2001).

The Federal Circuit’s approach appears to be inconsistent with the statute andcontrary to the policy concern expressed by the Supreme Court. The Court recognized thesocial value of a rigorous indefiniteness doctrine in a 1942 case, Justice Jackson stated:“A zone of uncertainty which enterprise and experimentation may enter only at the risk ofinfringement [suit] would discourage invention only a little less than unequivocalforeclosure of the field. Moreover, the claims must be reasonably clear-cut to enablecourts to determine whether novelty and invention are genuine.”17

As the Federal Circuit admitted, tolerating vague language at the Patent Officeyields difficult questions of claim interpretation. The Supreme Court anticipatedcontemporary claim construction problems when it warned: “Some persons seem tosuppose that a claim in a patent is like a nose�of wax which may be turned and twisted inany direction, by merely referring to the specification, so as to make it include somethingmore than, or something different from, what its words express.”18

Some ambiguity would not be too harmful if the public could rely on a predictablemethod of claim interpretation. Unfortunately, the Federal Circuit has not formulatedsuch a method. Recall that district court judges do a poor job of predicting FederalCircuit claim interpretation. Certainly, it follows that lawyers will have difficultycounseling potential infringers how an ambiguous claim term will be interpreted. Thediscussion in the advanced topic section below illustrates the difficulty of this

16 White v. Dunbar, 119 U.S. 47, 52 (U.S. 1886).17 United Carbon Co. v. Binney & Smith Co., 317 U.S. 228, 236 (1942).18 White v. Dunbar, 119 U.S. 47, 51 (U.S. 1886).


interpretative task.

[Insert advanced topic about here]

The Federal Circuit is keenly aware of the notice problems created by uncertainclaim construction. The law governing claim construction has long been in flux as courtshave searched for satisfactory methods. A significant development occurred ten years agoin the case Markman v. Westview Instruments.19 The case and its progeny made claimconstruction a question of law subject to de novo review in the Federal Circuit.Essentially, this means that the judges in the Federal Circuit have the power to choose themeaning of patent claims anew, ignoring lower court decisions. Alternative approacheswould have distributed power among fact-finders, trial court judges, and appellate judges.

Consolidating power in the appellate court offers the benefit of the experience ofthe Federal Circuit judges, but reliance on their experience reviewing patent cases has notpaid off. The goal of predictable claim interpretation has been thwarted by threeproblems. First, members of the Federal Circuit have been feuding over appropriatemethods of claim construction.20 One camp takes a very formal approach tointerpretation, and the other camp is more willing to rely on contextual clues to aidinterpretation. Second, the appellate court is too distant from the expert testimony andother facts that should be used in sensible claim construction. Judge Mayer of the FederalCircuit laments: “Because claim construction is treated as a matter of law chimericallydevoid of underlying factual determinations, there are no ‘facts’ on the record to preventparties from presenting claim construction one way in the trial court and in an entirelydifferent way in this court.”21 Third, in its formalist vein, the Federal Circuit has beenskeptical about the use of extrinsic evidence in claim construction. Extrinsic evidencecomes from outside the patent document and proceedings in the Patent Office. It includesexpert testimony, journal articles, dictionaries, and other outside evidence that mightreveal industry usage. Much of this extrinsic evidence is available to parties interested inmapping out the scope of a patent before they get near a courtroom. Potentially, greateruse of extrinsic evidence would strengthen the notice function, but it would also reducethe power of the Federal Circuit.22

Fuzzy boundaries are still possible even when claims are relatively precise andtheir literal scope is clear. Warner-Jenkinson Co. v. Hilton Davis Chemical Co.illustrates this point. The case featured yet another case of inadvertent infringement; theparties independently invented an improved process for purifying dye. Hilton Davisclaimed a version of the process operating at a pH ranging from 6.0 to 9.0. Warner-

19 116 S.Ct. 1384 (1996).20 Wagner and Pethebridge (2004). Both camps are relatively formal, and reject the use of policy

considerations during claim construction. In a district court ruling, Circuit Judge Richard Posner, sittingby designation, worried that a broad claim construction of a drug patent that could cover only traceamounts of the claimed compound would be contrary to patent policy. The Federal Circuit reversed,stating that claim construction is not a policy-driven inquiry, but rather a contextual interpretation oflanguage. SmithKline Beecham Corp. v. Apotex Corp., 365 F.3d 1306 (Fed. Cir. 2004).

21 Judge Mayer, dissenting, Lava Trading, Inc. v. Sonic Trading Management, LLC, et al., x F.3d x (Fed.Cir. 2006).

22 Burk and Lemley (2005) suggest that a great deal of scholarship about the textual interpretation ofcontracts points to the futility of relying too heavily on intrinsic evidence. Interpreting words can bedifficult. They report that recent Federal Circuit cases have had to decide plausible disagreements overthe meanings of the words “a,” “or,” “to,” “on,” “about,” “including,” and “through.”


Jenkinson’s technology operated at a pH of 5.0, outside the range of the relevant claim.Nevertheless, a jury found Warner-Jenkinson infringed under the doctrine ofequivalents.23

This doctrine expands the scope of a claim beyond the scope identified throughclaim construction.24 The original goal of the doctrine was to protect inventors against therisk that a pirate would spot a poorly drafted claim, introduce a slight variation in apatented technology that was outside of the claim, and escape literal infringement. Suchevasion could occur if the patent owner did not foresee the manner of imitation chosen bythe pirate, and therefore drafted its claim too narrowly. Today the doctrine applieswithout regard to the motives or methods of the alleged infringer – in particular, it appliesto independent inventors and those who make a good faith effort to invent around apatent, as well as to pirates.

The doctrine of equivalents corrodes the notice function of patents and increasesthe risk of inadvertent infringement. Warner-Jenkinson might have thought a pH of 5.0was outside the fence erected around processes with pH’s ranging from 6.0 to 9.0, butthey were wrong. The Supreme Court admitted: “[t]here can be no denying that thedoctrine of equivalents … conflicts with the definitional and public-notice functions ofthe statutory claiming requirement…,”25 nevertheless, they confirmed the availability ofthe doctrine against inadvertent infringers.

4 Public availability of boundary information

Determination of boundaries depends on publicly available information. Landdeeds are available in county registries. Most patent documents are also publiclyavailable, but inventors can delay release of important boundary information.

Many inventors act strategically to hide their claims from potential infringers.Inventors are allowed to draw out the patent application process for years if that servestheir interest — and sometimes it does. They often monitor the technology choices ofother firms and write their patent claims to cover the technology of potential licensees.The targeted firms may get locked into a technology choice and find themselves inunfavorable bargaining positions with the patent owner.26

In a notorious recent example, the firm Rambus participated in a semiconductorstandard setting organization while it Rambus was secretly pursuing a patent on the sametechnology. By participating in the standard setting process, Rambus learned informationthat it used to write claims that covered the standard. This strategy worked because thecompany added the claims to a patent that was pending when the standard setting process

23 After a round of appeals the case was sent back to the district court and it was dropped or settled. TheSupreme Court asked the lower courts to consider whether a doctrine called prosecution historyestoppel should have prevented Hilton Davis from using the doctrine of equivalents. Prosecution historyestoppel is one of several methods that courts use to screen cases and limit use of the doctrine ofequivalents. For more information about the doctrine of equivalents see Meurer and Nard (2005a).

24 Courts resolve questions about the doctrine of equivalents about 55% as often as they resolve questionsof literal direct infringement. www.patstats.org.

25 Warner-Jenkinson v. Hilton-Davis, 520 U.S. 17, 29 (1997).26 The strategic benefits of hidden claim drafting probably explain why the average prosecution time of

litigated patents is much longer than patents that are not litigated (Allison, Lemley, Moore, andTrunkey, 2004).


began. Thus, the “invention” was completed before the standard was set, but the claimswere written after the standard was set. Rambus dodged private suits alleging fraud andantitrust violation and was successful in enforcing its patent.27

It appears that hidden claims are becoming much more prevalent. One practiceused to keep claims hidden is to file “continuing” applications. Under US Patent Officerules (most other nations lack comparable procedures), once an original application isfiled, one or more continuing applications based on the same invention, but with differentclaims, can be filed. This gives the patent applicant additional opportunities to change theclaims over time, possibly catching unsuspecting innovators by surprise. The number ofcontinuing applications, shown as the heavy dashed line in Figure 1, has increased seven-fold since 1984 to about 120,000 per year (Quillen and Webster 2006). Continuingapplications now comprise about one third of all patent applications.

The patent system was reformed about ten years ago to reduce some problems ofhidden claims. The term of the patent was changed from seventeen years from issue totwenty years from application date, and patents are published after eighteen months,unless the applicant refrains from patenting outside the U.S. These reforms probably hada significant positive effect, but hidden claims are still a problem. Applicants can changeclaim language in patents without updating the published applications. The final claimlanguage is published only after the patent is issued and the gap between application dateand issuance is growing.28 Moreover, publication does little to prevent patent applicantsfrom introducing unanticipated new claims via continuing applications.

5 Possession and the Scope of Rights

In general, to own, one must possess. Although possession is not sufficient toguarantee ownership, the law by and large assigns initial property rights only to theextent that one possesses—exerts some control over—a property. A possessionrequirement means that the owner of a property is actually in a position to make thesocially desirable investments or transactions that are the key economic benefits ofproperty.

Simple possession rules also make it easier for third parties to recognize theexistence and scope of a property right. If farmer Jones owns and possesses a shirt, atractor, a house, and a farm, then neighbor Smith should know that he must askpermission before using that property. Furthermore, Smith probably won't have anytrouble figuring out that Jones is the owner.

The scope of most tangible property is easily described. Fences, surveys, and landrecords provide clear notice of the scope of rights to land. The physical structure ofgoods, machines, buildings, etc., defines those kinds of property — and gives crystalclear notice of the scope of rights. Rental, securitization, liens, and other legal devicesmay complicate the allocation of rights to use tangible property, but they are not relevantto what we mean by scope. Clear scope is present when a stranger recognizes when she isabout to use someone’s property and puts her on notice to seek whatever permissions are

27 The FTC recently ruled that Rambus violated antitrust law. That ruling is now on appeal. [Check inJune.]

28 That is, the time from initial application to patent issuance (Lemley and Moore, 2004).


necessary. Everyone recognizes that an unattended boat tied to a dock is someone’sproperty, and it should not be used without permission.

In some cases, when tangible property is taken from nature, the scope of theproperty rights is not so clear. In these cases, simple physical characteristics are not souseful for establishing legal boundaries because the relevant characteristics change overtime or are not fully known initially (that is, they are revealed over time). The miningdisputes discussed in Chapter 2 make this point. Another example comes from water law.In certain jurisdictions, the right to use water from a stream running through a propertydepends on the past consumption of others along the stream. Hence, a newcomer willneed to investigate her neighbors’ water use to determine whether and to what extentproperty rights already exist in the stream flow.

In the case of migratory wild animals, property law follows the “rule of capture”:you can own what you capture, but not the stock from which it came. Thus, whensomeone shoots a wild duck, she does not gain rights to the flock. It is easy to see howthe rule of capture promotes clear notice. Suppose the first hunter to shoot a duck in aflock actually gained ownership over the flock. It would be virtually impossible forhunters in the next county to recognize the flock was owned. Furthermore, the counter-factual property rule would invite endless disputes about who was the true owner of theflock, and which ducks belong to which flock.29

29 The rule of capture applies to natural oil and gas discoveries. A landowner owns the oil brought up from

Figure 1. The growth in original and continuing patent applications(Sources: Quillen and Webster 2006, NSF)

0

50,000

100,000

150,000

200,000

250,000

300,000

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

Pat

ent a

pplic

atio

ns

$0

$50,000

$100,000

$150,000

$200,000

$250,000

$300,000

R&

D (m

illio

n $1

992)

Original applicationsTotal US R&D Continuing applications


Similarly, the possession rule in patent law is designed to mitigate noticeproblems. The first paragraph of Section 112 of the patent statute requires that the patentdescribe how to make and use the invention in sufficient detail so that others can do so.This “enablement” requirement30 makes the patentee demonstrate the practical knowledgeneeded to usefully own the claimed invention.

This possession requirement allows courts to invalidate patent claims that are “toobroad” in the sense that the inventor did not really possess all the claimed technology. Afamous example concerns patents on the light bulb. Edison was not the first inventor ofthe incandescent light bulb. He had many competitors, and his light bulb built on manyearlier contributions.31 Sawyer and Man, were two competitors who obtained a light bulbpatent before Edison achieved his famous invention and they sued Edison. Their patentclaimed a light bulb with a “conductor of carbon, made of fibrous or textile material….”Edison made a light bulb with a bamboo filament that fell within the language of thebroad Sawyer and Man claim. The court ruled in favor of Edison because Sawyer andMan had actually only made a light bulb using carbonized paper as a filament. They didnot make light bulbs with other filaments drawn from the wide range of fibrous andtextile carbon based filaments, and in fact, most of those filaments would not work.Edison labored mightily to find a bamboo filament which worked very well—he triedover 6,000 different substances before settling on bamboo. But the Sawyer and Manpatent did not describe this important detail. They possessed the specific invention of alight using carbonized paper, but they did not possess the claimed knowledge to makeand use all “fibrous or textile” forms of carbon, including the bamboo later discovered byEdison. Therefore, the court invalidated Sawyer and Man’s claim because it claimedmore than they actually possessed – they claimed technology that was not yet invented.32

Ideally, enablement restricts patent scope so that inventors’ property rights do notstray far from the invention they actually possess. In the past, inventors had todemonstrate a working prototype or scale model of the invention in order to demonstratepossession. Inventors no longer need to provide a working prototype in order to obtain apatent, however, the general possession requirement remains central to patent law.

his wells, but the oil field itself is not owned nor does the landowner have any definite claim on futureoil withdrawals. Static physical boundaries do not work here because the underground structure of theoil field is not usually fully known until the field has been well-developed. For example, if two wellsthat are twenty miles apart both strike oil, it may not be clear whether they are tapping the sameunderground reserve or not. Giving ownership of the entire field to one well-owner will cause costlydisputes. Moreover, such a rule would inefficiently penalize the other well-owner if the fields are,indeed, separate. This would effectively discourage prospecting. Instead, the rule of capture works toreduce costly disputes and unearned ownership.

30 Some recent decisions have also read a “written description” requirement of possession into thisparagraph as well.

31 The Supreme Court noted other carbon filament, incandescent lamp inventions “in 1845, by Greenerand Staite in 1846, by Roberts in 1852, by Konn in 1872, by Kosloff in 1875, and by others.”

32 In patent law jargon, the broad claim was not enabled. Of course, the current statute was not in effect atthat time, however, a similar common law possession requirement existed. A modern counterpart can befound in Amgen, Inc. v. Chugai Pharmaceutical Co., 927 F.2d 1200 (Fed. Cir. 1991). Amgen invented arecombinant version of EPO, a human protein that stimulates red blood cell production, used to treatanemia. Chugai was the licensee of a different patent on purified natural EPO. Amgen claimed anyfunctional analog of the recombinant EPO that it actually made. This claim was broad enough to coverChugai’s EPO. After five years of experimentation Amgen was unable to identify which analogs hadthe desired properties. The claim was invalidated because of excessive breadth.


Thus, we are troubled by the many recent examples of patent claims that havebeen read broadly to cover infringing technologies that are distant from the inventionactually possessed by the patent owner. Many of these infringers have achievedsignificant inventions that they made independently of any help from the informationcontained in the patent. Consider, for example, the following two cases.

In a biotech case, Amgen obtained broad coverage from its patent claim on theprotein erythropoietin (EPO), a naturally-occurring hormone that promotes theproduction of red blood cells.33 EPO had been previously isolated and purified butattempts to obtain therapeutically useful quantities of the hormone from urine or bloodhad failed. Instead, Amgen isolated the human DNA that coded for EPO and inserted itinto Chinese hamster ovary cells. These genetically engineered cells produced humanEPO that could be efficiently isolated and purified. As a result, Amgen launched theblockbuster drug Procrit(tm) and they obtained a patent that claimed all “non-naturallyoccurring” EPO.

Later, an innovative biotech company, Transkaryotic Therapies, Inc. (TKT),developed an entirely different method of producing EPO. TKT did not insert humanDNA in a foreign host cell, but instead, they figured out a way to trick human cells intodirectly producing large amounts of EPO. Amgen successfully sued TKT (and its allyHoechst) for infringement.

Amgen faced two main obstacles en route to an infringement judgment. First, theclaim to all “non-naturally occurring” EPO had to be read broadly enough to cover theEPO made in human cells using TKT's technique. Second, this broad claim had to meetthe enablement requirement, that is, Amgen had to show it possessed the relevanttechnology. Amgen overcame these obstacles even though they clearly did not know howto perform TKT's technique. All that these two technologies had in common was thatthey shared the objective of producing large quantities of EPO outside of the humanbody.

A second case concerns the patent granted over relatively abstract ideas to Freenythat we discussed in the introduction. Freeny actually invented a kiosk for generatingaudio tapes and the like at retail stores. Nevertheless, because his patent claims werehighly abstract, they were interpreted to apply to a broad swath of e-commerce, eventhough nothing in his patent described general purpose transactions over the Internet.

Arguably, the Freeny and Amgen patents suffer from the same defect as the patentof Sawyer and Man: they do not demonstrate possession of the broad range oftechnologies that they claim. There are two clear problems with this policy. First, itmakes patent boundaries incredibly fluid over time. Third parties cannot rely on the plainmeaning of a term at the time of patent application – claim terms are allowed to changemeaning over time as technology advances. Second, it penalizes real innovators whooperate in the shadow of early, broad claims. Both of these factors tend to generate costlydisputes.

Many of the troubling cases in which patent claims are untethered from actualinventions arise in biotechnology. This is no accident. Several factors have contributed toa “flood” of biotech patents on very early-stage technologies: the Patent Office has

33 Amgen Inc. v. Hoechst Marion Roussel, 314 F.3d 1313. For more information see Feldman (2005).


granted thousands of patents on gene fragments, the Bayh-Dole Act has encouragedpatenting by academic researchers even though most of these inventions are very earlystage (Thursby & Thursby), and the Federal Circuit has changed patentability standardsto permit patents on technologies that are still far removed from practical application.34

Inevitably, a large number of patents on early-stage technologies lead to attempts bypatentees to assert these patents over later-developed technologies.

Patenting of early-stage technology is not the only situation where patents areasserted against later-developed technologies. For a variety of reasons, which weexplore separately in Chapter 9, abstract patent claims are particularly endemic tocomputer-related patents. Of particular importance, the law has changed to permit thepatenting of abstract software ideas, and the number of software patents granted hasincreased dramatically. The growth in the numbers of these patents also contributes to thefrequency with which patents are asserted over fundamentally more advanced later-developed technologies.

The increase in patents on early-stage technologies and the increase of patentswith abstract claims has lead to lawsuits against later-developed technologies with greaterfrequency. This has increased the importance of clear possession rules. But judges havenot consistently limited ownership to the technologies actually possessed, perhapsbecause some judges have felt an impulse to reward “pioneer” inventors. But the courthas not been uniform about this, with some judges developing controversial new rules tolimit the reach of patents over later-developed technologies. Legal scholar RobinFeldman (2005) documents the state of tumult in patent law concerning this issue forbiotech patents.35 Unfortunately, we suspect that the net effect of both the expansivereading of patent rights and the new corrective measures has been to increase uncertaintyand costly disputes.

6 The Patent Flood

A large number of property rights held by many different owners can make theclearance of rights for new investment costly. If clearance costs grow too large, thencomplete clearance becomes infeasible, and firms will only do a cursory clearance or,perhaps, none at all.

In Chapter 2, we discussed how property law uses various rules to limit

34 In Brenner v. Manson, 383 U.S. 519 (1966), the Supreme Court held that “[A] patent is not a huntinglicense” and early-stage results that did not yet demonstrate a practical, useful application were notpatentable. But in In re Brana,51 F.3d 1560 (1995), the Federal Circuit held that the burden of provinga lack of useful application fell on the Patent Office; the court determined that a drug compound withpromising test results against tumors in mice had sufficient evidence of practical utility. The FederalCircuit appears to be heading back into compliance with Brenner. In In re Fisher, the court ruled that agene fragment patent failed to meet the utility requirement.

35 Judges have recently developed new doctrines for interpreting the enablement and written descriptionrequirements found in Section 112, see for example, Regents of Univ. of Cal. v. Eli Lilly & Co., 119F.3d 1559 and Plant Genetic Sys. N.V. v. DeKalb Genetics Corp., 315 F.3d 1335, but these legalinnovations have been controversial. See Judge Bryson’s concurrence in Chiron Corp. v. Genentech,Inc., 363 F.3d 1261 and Judge Rader’s dissent in LizardTech, Inc. v. Earth Res. Mapping, Inc., 433 F.3d1376, even among the various judges of the Federal Circuit. Consequently, these doctrines have notbeen uniformly applied, possibly leading to increased uncertainty about patent boundaries.


fragmentation of rights, including primogeniture, limits on future interests, property taxesand registration requirements, and zoning and subdivision restrictions.

Patent law also has rules that limit the proliferation of rights. First, an inventormust demonstrate an invention has sufficient practical utility in order to receive a patent.Also, patents are not granted for inventions that are obvious improvements on previoustechnology. These patentability requirements work to weed out the potential clutter ofpatents on inchoate or minor inventions. In addition, patent renewal fees work to weedout less valuable patents.

However, the evidence suggests that over the last two decades the number ofpatent rights has proliferated dramatically. The number of patent applications more thantripled from 1980 (104,329) to 2004 (356,943). That would be great news if itrepresented dramatic growth in American innovation but the story is not so simple. Thegrowth was not caused mainly by a dramatic jump in R&D spending. Instead, the numberof original patent applications has outstripped the growth of US R&D spending (seeFigure 1), especially since the mid-1990s.

Empirical research by Bronwyn Hall finds that patent application growth datadisplays “a very significant structural break between 1983 and 1984” which she attributesto the pro-patent policies of the Federal Circuit. Indeed, Henry and Turner (2006) identifya pro-patentee structural shift in court rulings that occurred about this time. Furthermore,Hall (2005) finds the growth comes from U.S. firms in the “electric machinery,electronics, instruments, computers, and communication equipment” industries. Much ofthis growth took place during the 1990s and may be associated with subject matterexpansion for software-related inventions(Bessen and Hunt 2007).

The growth in patent applications and grants has been accompanied bycomparable growth in the number of claims per patent. Even though each patent issupposed to protect only one invention, patents may have multiple claims, many patentshave dozens of claims. Inventors write claims to protect different aspects of an invention,for example by claiming a product and also processes for making and using the product.Furthermore, they write claims of varied coverage to hedge against the risk that certainclaims are invalidated or read narrowly. Allison and Lemley (2002) compared patentsfrom the mid 1970s to patents from the mid 1990s and found that the mean number ofclaims had grown from 9.94 to 14.87, nearly a fifty percent increase.

The growth in the numbers of patents and the growth in the numbers of claims donot necessarily imply a comparable increase in search costs associated with clearance. If,for example, patents could be neatly divided into technology classes so that only a smallclass needs to be searched, then search costs might remain reasonable despite the hugeincrease in numbers. But evidence from litigation suggests that this is not so. Bessen andMeurer (2005) found that about a quarter of all lawsuits between public firms involvedfirms that patented in very different technology classes and which were in unrelatedindustries. Indeed, it is not hard to find examples where firms are sued over patentscovering distant technologies. For example, Amazon.com has been sued by a firm with apatent covering cable TV movie selection and another firm with a patent on a bank ATMinterface.36 This implies that a complete clearance procedure requires careful examination

36 Statement of Paul Misener before the Subcommittee on Courts, the Internet, and Intellectual Property,Committee on the Judiciary, United States House of Representatives, June 15, 2006.


of very large numbers of patents indeed. As noted in the Introduction, David Martin ofM-Cam estimates that a typical web shop may need to check over 11,000 patents.Clearly, costs of clearance can be substantial.37

Moreover, the law actually imposes penalties that discourage thorough clearance.Knowledge gained through a patent search increases the risk of a finding of willfulness ina patent lawsuit. Firms are discouraged from reading patents because of fear of thefinancial penalties that might follow a finding of willfulness. The disincentive is greatestwhen a firm believes there are a large number of weak patents in the technologicalvicinity of its innovation. If the firm is unlikely to license the patents, then reading themsimply increases the risk of a finding of willfulness. On the other hand, when a firmbelieves there are a small number of strong patents nearby, it would read and licensethem, or be careful to invent around them.38

The cost of clearance ratchets up even more when patents have fuzzy boundariesand when many patents are likely to be found to be invalid. With these uncertainties, atechnology investor will have to examine many patents of vague scope and dubiousvalidity. Note that even though the Federal Circuit has strengthened patent rights bymaking changes to the law that reduced the probability of invalidity, the share of districtcourts decisions that invalidated a patent is still 27%.39 This introduces an element of riskinto clearance decisions and it disrupts attempts to invent around patents. Kodak relied onCarr’s opinion that several Polaroid patents were invalid. If they thought those patentswere likely to be valid, they would have chosen another product design, or perhaps stayedout of the market.

There is abundant evidence that many technology firms follow the same path asRIM and invest little in patent search and clearance. A recent survey of the members ofthe Intellectual Property Owners organization found that 65 percent of respondentsdisagreed with the statement: “We always do a patent search before initiating any R&Dor product development effort.”(Cockburn and Henderson, 2003)

Finally, in addition to the costs of searching and identifying rights to clear, atechnology investor may also face substantial costs of transacting for those rights.40

37The semiconductor and communication industries also face remarkable clearance problems. In 2002,Intel Vice President and Assistant General Counsel Peter Detkin claimed that there were about 90,000 CPUpatents held by over 10,000 parties, and 420,000 semi/system patents held by over 40,000 parties. T Also,the organization developing a standard for the third generation of wireless telephones issued a call fordisclosure of relevant patents and offers of licensing terms. In reply, they received notice of 7600 patentsand a combined royalty request summing to 125% of sales revenue (Lemley and Shapiro, 2007).38 Unless the firm intends to pirate the patented technology.39 Henry and Turner (2005) studied reported patent opinions from 1953-2002. They selected opinions

containing substantive holdings and placed them into three categories: invalid, valid and not infringed,and valid and infringed. The 27% figure is the fraction of substantive holdings that resulted ininvalidity. The fraction in the pre-Federal Circuit era was 55%.

40 When different firms own patents that are complements non-cooperative pricing drives total licensefees above the price that a monopolist would charge. Real estate developers are familiar with anotherbargaining problem created by the need to acquire a large number of property rights – the hold-upproblem. If several specific parcels of land must be acquired before a development project is feasible,then a developer would worry that one of the land owners might hold-out, and refuse to sell her landunless she gets a substantial premium over the market price for the land. When land owners first challenged plane flights over their land as trespass, the courts refused to enforcethe traditional scope of property rights that extend out to space. The scope of property was sensibly


Semiconductor manufacturing is one industry where this problem is felt acutely.“Semiconductor manufacturing is … notoriously complex, integrating an array of processand product technologies that cover aspects of the circuitry design, materials used toachieve a certain outcome, and methods used in the wafer fabrication process. As severalindustry representatives pointed out, a given semiconductor product (say, a new memoryor logic device) will often embody hundreds if not thousands of ‘potentially patentable’technologies that could be owned by suppliers, manufacturers in other industries, rivals,design firms, or independent inventors.”(Hall and Ziedonis, 2001) RosemaryZiedonis(2005) finds that fragmentation is such a serious impediment to licensing that itdrives patenting behavior by certain firms in the semiconductor industry.41

* * *

In this chapter we have uncovered four aspects of the notice problem. To begin,inventors can hide patent claims, and thus boundary information, from the public. Next,even when the relevant public has access to the patent claims they are very difficult tointerpret. Even assuming the claims are available and clear, there is a danger that themeaning of claim language will change (and become broader) over time. Finally, evenwhen claims are available, clear, and fixed over time, the cost of searching for relevantpatents can be quite high. High search costs arise directly because of the high number ofpatents that potentially apply to certain technologies, and indirectly because of the highrate of invalidity and the willfulness doctrine both discourage search. In combination,these four problems can reinforce each other, e.g., a large number of fuzzy orunpredictable patents make clearance procedures especially fruitless.

The law and institutions that support notice in the patent system fall far shortcompared to similar institutions supporting notice for tangible property rights. It is hardlysurprising that land title insurance is cheap and readily available while patentinfringement insurance is limited, expensive and not widely offered. Nor is it surprisingthat inadvertent infringement is a serious problem for patents but not for tangibleproperty. In the next chapter we begin our examination of empirical evidence on theperformance of patents as a property system. This may reveal the economic significanceof patent notice problems.

limited to avoid the great social cost that would have arisen as airplane owners tried to obtain flyoverrights.

41 Ziedonis studied a sample of 67 semiconductor firms from 1980–1994, and found “that capital-intensive firms patent more than five times as aggressively in response to average levels offragmentation in markets for technology as firms of average capital-intensity, even controlling fordifferences in R&D spending and size. Moreover, and extending the earlier findings of Hall andZiedonis (2001), I find that capital-intensive firms do not patent more intensively (again, controlling forother factors) unless they build on fragmented pools of outside technologies. There is little evidence tosuggest that these findings are explained by underlying shifts in technological opportunity, divergentR&D efficiencies, or other unobservable sources of heterogeneity within the sample.”


Advanced Topic: Claim ConstructionGood patent policy analysis is not possible unless one has a basic understanding

of patent claims, how they are interpreted, and how they create property rights. So, inthis section, we present the highlights from a patent lawsuit involving a claim relating toa simple invention. Even though the invention is simple, you will soon see thatunderstanding and applying the claims can be quite difficult even for experienced patentjudges.

U.S. Patent No. 4,018,260 relates to a home improvement invention, specifically a set ofborder pieces used to attach a fabric wall covering to a wall. Drawings from the patentare displayed on the left hand side of Figure A1. The patent suggests the advantage of theinvention is that it makes it easier for an inexperienced person to hang wall covering. Therelevant portion of the disputed claim 1 reads as follows:

1. An assembly of border pieces for creating a frameworkattachable to a wall or other flat surface for mounting afabric sheet which is cut to dimensions at least sufficientto cover the surface, said assembly comprising

(a.) linear border pieces and

(b.) right angle corner border pieces which are arranged inend-to-end relation to define a framework that follows theperimeter of the area to be covered…

The patentee, Unique Concepts, sued Brown for patent infringement. Thedefendant made and sold a set of border pieces for hanging wall covering. The outcomeof the case turned on the interpretation of claim 1, and in particular of the italicizedlanguage: right angle corner border pieces, labeled 15 in the figure.42 The claim alsodescribes linear border pieces (labeled 14) and distinguishes them from the right anglecorner border pieces. Brown’s set of border pieces contained trapezoidal pieces with aforty-five degree cut (labeled 4 on the right-hand side of Figure A1). Two trapezoidalpieces could be connected to serve the same function as the right angle corner border

42 Unique Concepts, Inc. v. Brown, 939 F.2d 1558 (Fed. Cir. 1991).

Figure A1. Patent Claim and Alleged Infringing Alternativ


pieces. Brown did not use corner pieces as in the patent. Thus, the court needed to decidewhether the scope of claim 1 was broad enough to include the defendant’s product.

The majority of judges ruled there was no infringement, but Judge Rich dissentedand offered an alternative claim construction that was broad enough to support a findingof infringement. In effect, Judge Rich argued that the term “right angle corner borderpieces” is not restricted to pre-formed corner pieces and it includes the allegedinfringer’s configuration. The disagreement concerned not only the language in theclaims themselves, but also what the patentee’s original language claimed, how this waschanged in response to Patent Office objections, and how the invention was describedelsewhere in the patent.

This dispute reveals the troubling indeterminacy of claim construction. We suspect theuncertainty about claim boundaries is even greater for more complex technologies,abstract claim terms and early-stage technologies.

if you can't tell the boundaries, then it ain't property - research on

Documents